autonomous movement of the enzyme-filled nanomotors we analyzed their behavior in the presence of hydrogen peroxide and glucose at different concentrations. We used nanoparticle-tracking analysis (NTA), a technique complementary to DLS that uses laser light scattering in combination with a charge-coupled camera (CCD) and a microscope, to
provide individual particle-by-particle analysis of colloidal particles instead of an assemble size distribution as shown by DLS (see section 5.2.9). The Stokes-Einstein equation is then used to determine the size of the structures by correlating the tracking coordinates from the Brownian movement to the particle size as shown in our previous study on platinum-driven nanomotors.11 In this equation the hydrodynamic diameter of the supramolecular nanomotor
d is inversely related to the time-dependent particle diffusion coefficient D(t), which
however is valid only when no fuel is present in the system and the particles move under Brownian motion (D(t)=TKB/3πηd, with KB the Boltzmann constant, η the viscosity, and T
the temperature). The fast directional autonomous movement of the nanomotors in the presence of the fuel makes their sizes to “appear” smaller compared to the same structures before adding the fuel, due to the inverse relation between diffusion and (apparent) size. Since the technique provides additional visualization of the particles, it is also suitable, as we showed previously, for tracking the non-Brownian motion when fuel is added to the self- assembled structures.
To test the expected directional movement in our enzyme-driven nanomotors and make sure that the fuel addition was not responsible for the change in the size, we investigated the effect of the addition of hydrogen peroxide and glucose to empty stomatocytes. As expected, no change in their Brownian motion and trajectories was observed (see supplementary movie 1, supplementary figure 11). However, addition of hydrogen peroxide of different concentrations to the catalase-filled stomatocytes solution resulted in a clear shift in their apparent sizes to smaller values compared to the same structures in the absence of fuel (see supplementary figure 13). Additionally, a clear change of their trajectories from a non-directional Brownian motion to a propulsive directional movement was observed (see supplementary figure 12). When the fuel was fully consumed, their original size was measured again by NTA, demonstrating that the effect was due to the propulsive movement of the nanomotors. Furthermore, the addition of hydrogen peroxide to a mixture of 90 % empty stomatocytes and 10% catalase nanomotors (v/v) showed simultaneously the autonomous directional movement of the nanomotors and the expected Brownian motion of the empty stomatocytes (see supplementary movie 2, note the fast tumble and run movement of the nanomotors). This experiment further confirms that the movement of the assembled nanomotors is autonomous and is not caused or affected by any drift or flow within the chamber, which is only expected at much higher fuel concentrations than used in our system, due to the fast accumulation of gases within the chamber. As shown in our previous report on platinum-driven nanomotors the ability of the NTA technique to
measure the trajectories and x,y coordinates of the single particle nanomotors allowed for a closer analysis of their movements by studying their paths and their average mean square displacements (MSD).11, 42 We used the self-diffusiophoretic model proposed by Golestanian and coworkers to determine the speed of the nanomotors.42 The model indicates that the directional movement of micron size Janus sphere motors is the result of both rotational and translational diffusion. The model has two limiting forms, a parabolic component for short periods of observation and a linear component for long periods. The fitting of the experimental MSD data of our enzyme-driven nanomotors allowed only for the observation of the parabolic component. This was due to the limitations of the nanosight system in the movement analysis of nanometer scale objects that prevented the tracking of the nanomotors for long periods of time and at high capture rates. Both the trajectories and the average MSD’s of 105 nanomotors at 3 hydrogen peroxide concentrations (11 mM, 50 mM and 111 mM) were measured and the propulsive and directional movement of the nanomotors was determined from the fitting of the parabolic fit of the MSD dependency in time according to the equation <r2> = 4Dt + (vt)2 (see figure 5a and supplementary figure 14) with D, the diffusion coefficient and v – speed of the nanomotors. The movement of the nanomotors without fuel (controls) showed only a linear <r2> = 4Dt dependency typical for a Brownian motion. The average speeds of the nanomotors at these concentrations were found to be 15 μm/s, 26 μm/s and 60 μm/s (see figure 5a). The bio-hybrid catalase-driven nanomotor therefore runs at remarkable high speeds of 176 body lengths/s in 100 mM hydrogen peroxide concentrations, which is 3 times higher than the speed of our previously reported platinum-driven nanomotors.11 This high efficiency is most probably due to the combination of high catalytic activity of the catalase molecules and the excellent encapsulation efficiency of the enzymes compared to the stomatocytes filled with the catalytically active platinum nanoparticles. We also think this is due to the special design of our nanomotor system, which confines the enzymes in a small compartment with a very small opening which forces the gases to be expelled through a nanometer pore. This design is much different from the traditional Janus particles where the substrates are released from a larger surface.
Figure 5: Movement analysis of the one and two-enzyme-driven supramolecular nanomotors. a) The velocity of catalase-filled stomatocytes at different fuel concentrations; The velocity was extracted from the fitting of the average MSD of the catalase-filled stomatocytes at different concentrations (11-111 mM H2O2), calculated from the tracking coordinates of on average 105 particles (a bigger version of the MSD curves can be found in supplementary figure 14). b) The velocity of GOx-catalase two enzyme-driven nanomotors at different fuel concentrations; The velocity was extracted from the fitting of the average MSD at different concentrations (5 and 10 mM glucose), calculated from the tracking coordinates of on average 100 particles (a bigger version of the MSD curves can be found in supplementary figure 15). c) Schematic representation of the size dependent inhibition and protecting effect of the stomatocytes in GOx-catalase two enzyme-driven nanomotors. Small inhibitor sodium azide is able to diffuse inside of the nanomotors deactivating the enzyme while large proteases are not able to get in. d) MSD of GOx-catalase two enzyme-driven nanomotors in the presence of catalase or trypsin added externally to the mixture. The velocity was extracted from the fitting of the average MSD of the curves (a bigger version of the MSD curves can be found in supplementary figure 16). No change in the speed of the nanomotors is detected.
We subsequently tested the stomatocytes containing the two-enzyme cascade system based on glucose oxidase and catalase with glucose as a fuel (see figure 5b). The ratio between catalase and GOx was selected to be 1:3 (w/w), taking into account the known difference in activities of the two enzymes. The GOx-catalase nanomotor was observed to become more active in time and increased its speed several seconds after the addition of the glucose. We attribute this behaviour to the slower GOx enzyme, which requires oxygen to start the catalytic process. For this reason, we used aerated MilliQ water to perform the rest of the experiments. The movements of particles at two glucose concentrations are shown in figure 5b and supplementary figure 15. As can be seen the two-enzyme nanomotor is able to propel at very low concentrations of glucose, even down to 5 mM. This is a much lower value when compared to a previously reported example where carbon nanotubes were used, to which the same catalytic enzyme combination was attached. In that case a much higher concentration of glucose of 100 mM was required.29 We think this is because the reaction in the stomatocytes is concerted in the nano-cavity and thus faster transfer of substrates between different enzymes occurs. This is not the case if the enzymes are chemically attached to the surface of the motors, as transfer of substrates relies on slow diffusion in solution.
In a final series of experiments, we investigated whether the motion of the supramolecular nanomotors could be manipulated by controlling the activity of the entrapped enzymes. Sodium azide is a known small inhibitor of catalase. Its anion binds to the heme iron center in the active site of the enzyme. As expected, the addition of sodium azide irreversibly inhibited the decomposition of hydrogen peroxide and consequently the production of the propelling oxygen gas necessary for the functioning of the stomatocyte motor. After the addition of the inhibitor, both the trajectories of the nanomotors and their sizes indicated the recovery of the Brownian motion characteristics of the nanomotors in the absence of fuel. The inhibition of the catalase inside of the stomatocytes was possible due to the small size of the sodium azide, which was able to diffuse inside of the stomach. In the case the inhibitor was a proteolytic enzyme, for instance trypsin, its larger size should prevent it from diffusing inside of the nanomotors to inhibit the activity of the enzyme (see figure 5 c, d). To test the protecting effect provided by the stomatocyte we exposed the GOx- Cat nanomotors to 434 μM trypsin and analyzed the movement of the nanomotors after protein addition (see figure 5d). Both enzymes (GOx and CAT) were able to work in cascade inside the stomatocytes to produce the propelling gas, therefore the presence of the proteolytic enzyme did not have any noticeable effect on the function as nanomotors (see
figure 5c, d). A small decrease in the speed of the nanomotors was observed most probably due to the decrease in the absolute system concentration (fuel and particles concentration) upon protein addition. The encapsulation of the enzymes inside the stomatocytes is of great importance as it provides protection against deactivating elements present in biological environments, such as proteases. The nanomotor design offers therefore a clear advantage compared to other enzyme nanomotors, especially when applying these nanomotors in biologically related applications due to their high efficiency and activity at very low concentrations of naturally occurring fuels.
4. Conclusions
In summary, we have developed a strategy to incorporate sensitive proteins or enzymes with very high encapsulation efficiencies inside the cavity of polymeric stomatocytes via a process of shape transformation of polymersomes under mild conditions, while fully retaining their activity. The encapsulation of the two enzymes GOx and catalase allows the stomatocytes to propel as nanomotors using glucose as an alternative fuel for hydrogen peroxide at biologically relevant concentrations, i.e. at only 5 mM. This efficiency is probably attributed to the compartmentalization and confinement of the enzymes in such a nano-vector. The morphology of these nanomotors provides protection of the enzymes within their cavities from proteolytic enzymes that are available in a biological environment thus providing a broader scope to the nanomotor design for biological applications, e.g. in living cells. Besides its application for nanomotor assembly, this strategy of encapsulation, release and protection of proteins within a nano-vesicle containing a large pore (stomatocyte) could be useful to other fields such as drug/protein delivery or nanoreactor applications. When the nanomotors are further optimized with respect to control in movement and directionality, they could be useful for applications such as biosensing, protein & DNA isolation and detection or immunoassays. Nanomotors could rapidly in situ recognize, isolate and enrich target biomolecules, such as DNA, proteins and cells, in untreated biological samples. Our nanomotor assembly and the strategy of encapsulation provide high flexibility in the cargo-load and hold therefore considerable potential for future research in the biomedical field.
5. Experimental
5.1 Experimental materials and instruments
All chemicals and enzymes were used as received unless otherwise stated. For the block copolymer synthesis, styrene was distilled before use to remove the inhibitor. Anisole and N, N, N’, N’’, N’’-pentamethyl-diethylenetriamine (PMDETA) were purchased from Sigma Aldrich. Ultra pure MilliQ water, obtained with the help of a Labconco Water Pro PS purification system (18.2 MΩ), was used for the procedures of polymersome self-assembly and the dialysis experiments. Dialysis membranes MWCO 12-14000 g mol-1 Spectra/Por® were used where required. Ultrafree-MC centrifugal filters 0.22 μm were purchased from Millipore. Sodium nitrate was purchased from Merck. Catalase (E.C. 1.11.16) from Bovine Liver, lyophilized powder 2000-5000 U mg-1 was purchased from Sigma Aldrich. Glucose Oxidase (E.C. 1.1.3.4) from Aspergillus niger Type II lyophilized powder 228.25 U mg-1 was obtained from Sigma Aldrich. Peroxidase from Horseradish (E.C. 1.11.1.7) Type I, 50- 150 U mg-1 solid and Ampliflu™ Red were purchased from Sigma Aldrich.
Proton nuclear magnetic resonance spectroscopy: 1HNMR spectra were recorded on a Varian Inova 400 spectrometer with CDCl3 as a solvent and TMS as internal standard.
Gel permeation chromatography: A Shimadzu Prominence GPC system equipped with a
PL gel 5 μm mixed D column (Polymer Laboratories) and differential refractive index and UV (254 nm) detectors was used. THF was used as an eluent with a flow rate of 1 mL min-
1. Polystyrene standards in the range of 580 to 377,400 g mol-1 were used for calibration.
Transmission electron microscopy: TEM experiments were performed on a JEOL 1010
microscope equipped with a CCD camera at an acceleration voltage of 60 kV. Samples were prepared by placing 5 µL of the solution on a carbon-coated Cu grid (200 mesh, EM science) and they were allowed to air-dry for at least 24 hours. Processing and analysis of TEM images was performed with ImageJ, a program developed by the NIH and available as public domain software at http://rsbweb.nih.gov/ij/.
Cryogenic transmission microscopy: The cryogenic transmission electron microscopy
(Cryo-TEM) experiments were performed on a JEOL TEM 2100 microscope (JEOL, Japan) and processed and analysed with ImageJ. Samples for CryoTEM measurements were prepared by first treating the grids (Quantifoil R2/2 Cu 200 mesh grids) in a 208HR sputter coater for 20 seconds. Afterwards, 3 µL of sample was brought on the grid and blotted in a
FEI Vitrobot Mark IV, at 100 % humidity. Subsequently, the grid was blotted and directly frozen in liquid ethane.
AF4-UV-MALS-QELS: The asymmetric flow field flow fractionation – UV – QELS (AF4-
UV-QELS) experiments were performed on a Wyatt Eclipse AF4 instrument connected to a Shimadzu LC-20A Prominence system with Shimadzu CTO20A injector. The AF4 was further connected to the following detectors: a Shimadzu SPD20A UV detector, a Wyatt DAWN HELEOS II light scattering detector (MALS) installed at different angles (12.9 º, 20.6 º, 29.6 º, 37.4 º, 44.8 º, 53.0 º, 61.1 º, 70.1 º, 80.1 º, 90.0 º, 99.9 º, 109.9 º, 120.1 º, 130.5 º, 149.1 º, and 157.8 º) using a laser operating at 664.5 nm, a Wyatt Optilab Rex refractive index detector and a QELS detector installed at an angle of 140.1o. Detectors were normalized using Bovine Serum Albumin and for the enzyme molecular weight calculations, dn/dc of 0.1850 was used. The AF4 channel was pre-washed with running solution of 5mM NaNO3, which was also used for the separation. The processing and analysis of the LS data
and hydrodynamic radii calculations were performed using Astra 6.1.1. All AF4 separations were performed on an AF4 short channel with regenerated cellulose (RC) 10 kDa membrane (Millipore) and spacer of 350 µm.
Nano-particle tracking analysis: These experiments were performed on a Nanosight
LM10HS instrument equipped with an Electron Multiplication Charge Coupled Device (EMCCD) camera. This camera was mounted on an optical microscope in order to track the light scattered by the injected particles that are present in the focus of the 80 μm beam generated by a single mode laser diode with a 60 mW blue laser illumination (405 nm). The solution containing polymeric vesicles was adjusted to a concentration between 107 and 109 particles mL-1. Solutions were injected into the nanosight sample chamber. The chamber has a 0.5 mL size from which a volume of 120x80x20 micrometer was visualized under the microscope. The Brownian motion of the nanoparticles was tracked with 30 frames s-1. NTA 2.2 software was used to track the vesicles and from this, their mean square displacement (MSD) was calculated. The method of MSD calculation was explained in detail elsewhere.11
Fluorescence measurements: The fluorescence measurements were performed on 96-
black well F-Bottom microplates (Greiner Bio-One) on a Berthold TriStar² LB 942 Multidetection Microplate Reader equipped with a 550 nm excitation filter and a 610 nm emission filter.
5.2 Experimental procedures
5.2.1 Synthesis of poly(ethylene glycol)44-b-poly(styrene)167: Amphiphilic block copolymer PEG-b-PSwas synthesized according to literature procedures using atom-transfer living radical polymerization.43 Briefly, to a dry Schlenk tube equipped with a stirring bar, copper bromide (CuBr) (45 mg, 0.32 mmol) was added under argon atmosphere. The Schlenk tube was then sealed with a septum, evacuated for 15 min after which argon was filled back into the flask. PMDETA (66 µL, 0.32 mmol) was dissolved in 0.5 mL of anisole and added into the CuBr. The mixture was left stirring for 15 min with argon bubbling through the solution for oxygen removal. Subsequently, poly(ethylene glycol) macro initiator (215 mg, 0.10 mmol) was dissolved in 1 mL of anisole and added into the Schlenk tube. The Schlenk tube was inserted in an ice bath, and the solution was degassed for 15 min. Afterwards, distilled styrene (5 mL, 43.6 mmol) was inserted into the Schlenk tube. The mixture was degassed by three freeze thaw cycles after which the Schlenk tube was inserted into a preheated 70 ºC oil bath overnight. Dichloromethane (CH2Cl2) (75 mL) was
then added into the polymer solution and the organic layer was extracted with aqueous 65 mM EDTA solution (3 × 150 mL). The aqueous phase was washed with CH2Cl2 and the
organic layers were combined and dried with MgSO4. The solution was then concentrated
and the polymer was precipitated in cold MeOH, filtered and dried overnight. The amphiphilic polymer obtained, PEG44-b-PS167 had a number average molecular weight (Mn)
of 19.6 kgmol-1 and a Ð of 1.07.
5.2.2 Preparation of polymersomes for either solvent addition or reverse dialysis shape transformation methods: Block-copolymer PEG44-b-PS167 (20 mg) was dissolved in 2 mL
of a mixture of distilled THF and dioxane (4:1 v/v) in a 15 mL vial equipped with a magnetic stirring bar. The vial was capped with a rubber septum. MilliQ (3 mL) was added via a syringe pump with a rate of 1 mL h-1 while stirring the solution vigorously. The resulting cloudy suspension was transferred into a dialysis membrane (SpectraPor, molecular weight cutoff: 12,000-14,000 Da, flat width 25 mm), which was prior swollen in MilliQ for about 30 min. The polymersomes were dialyzed against water (1000 mL) for at least 24 hours.
5.2.3 Stomatocyte formation via the reverse dialysis of polymersomes: Rigid membrane
PEG44-b-PS167 polymersomes (700 μL, 10 mg mL-1) were placed in a dialysis membrane
(SpectraPor, molecular weight cut-off: 12,000-14,000 Da, flat width 10 mm), which was swollen in MilliQ for 30 min. The dialysis solution was composed of 150 mL MilliQ (50% in volume) and a mixture of 120.0 mL THF and 30.0 mL dioxane (4:1 v/v). To monitor the
shape transformation by cryogenic transmission microscopy, samples were withdrawn from the dialysis bag at 30 min, 60 min., 90 min., 120 min., 150 min. and 180 min.
5.2.4 Stomatocyte formation via the solvent addition method: An aqueous rigid
polymersome solution based on PEG44-b-PS167 block copolymer (500 μL, 10 mg mL-1) was
transferred into five 5 mL vials equipped with a magnetic stirring bar and a septum. A